Research On Node Localization Algorithm Based On DV-Hop For Wireless Sensor Networks

Localization is one of the core technologies in wireless sensor networks, it is the foundation of some practical applications such as target monitoring, target identification and target tracking and so on. It is impossible to make all the sensor nodes deployed in the sensing area as the special nature of the application environment, and also it is not possible to make the sensor nodes apply the GPS or other external devices to position in result of the low cost and low energy consumption of the sensor nodes. To guarantee the low error, low energy consumption and the high efficiency of the localization algorithm, the study on the positioning scheme of the sensor nodes became essential in wireless sensor networks.At present, the localization algorithms are divided into range-free and range-based scheme based on whether need to measure the distance of the nodes, the range-based localization algorithms need to measure the real distance of the nodes by its own devices or other additional devices, which makes its higher hardware requirement, so it is not easy to implement and save cost of the nodes. The main range-based localization algorithms contains: RSSI(Received Signal Strength Indicator), TOA(Time of Arrival), TDOA(Time Difference of Arrival), AOA(Angle of Arrival) and so on. There is no need extra devices to measure the distance or the angel information in range-free localization algorithms, it can finish the position of the unknown nodes by the topology of the networks. Now, many range-free localization algorithms and the improved algorithms have been proposed, such as:Centroid algorithm, DV-Hop(Distance Vector Hop) algorithm, APIT(Approximate Perfect Point-In-Triangulation Test) algorithm, Amorphous algorithm and so on. the most typical algorithm is DV-Hop, which owns the features of simple, easy to implement, low complexity, low hardware requirements and so on. The study of the DV-Hop of range-free localization algorithms plays a great significance.A whole study of the idea of DV-Hop localization algorithm was carried out in this article. DV-Hop algorithm is a range-free algorithm, the localization of the unknown nodes don't need measure the real distance by extra devices, the specific process follows:firstly, anchor nodes broadcast the information, the unknown nodes and the anchor nodes get the hops to other anchor nodes; secondly, anchor nodes calculate the average hop distance and send it to its neighbor nodes, the unknown nodes calculate the estimate distance to the anchor nodes using the product of the hops and the average hop distance; lastly, after achieving the distance to all the anchor nodes, the unknown nodes calculate its coordinate by the maximum likelihood estimate method. Although DV-Hop algorithm is a distributed localization algorithm without measuring, but it's flaws and shortcomings also exist. At first, this article analyzed the process of the DV-Hop algorithm, listed the proved main source of error from the aspects of external factors and internal factors, for example:the exist of the bad nodes of the networks, the shortage of the mathematical methods used by the algorithms and so on. Then through the simulation and serious analysis, a new main source of error was proposed. It is the truth:the different position and different number of the anchor nodes play an important on the positioning error, which is proved to be correct by a series of simulations. The main reason is the generations of the accumulated error of the average hop distance, when an anchor node is farther, the hops to the unknown nodes is also more, so the error is also higher. When the algorithm is used in the large-scale networks, the accumulated error will be more evidence.Aimed at the error resource summarized in this article, we proposed an improved algorithm based on optimal anchor nodes. The idea of the improved algorithm is choosing the anchor nodes which can make the good positioning precision to be a reference anchor node, we can set a threshold for every node, every node select only a certain number of anchor nodes and a certain region anchor nodes to achieve the position rather than select all the anchor nodes. The results of large number of simulation experiments showed that the choice of the number of the optimal anchor nodes was related to the following aspects:the size of the sensing area, the average connectivity of the unknown nodes and the anchor nodes, the communication radius of sensor nodes and the proportion of the anchor nodes. We gave the select criteria of the number of the optimal anchor nodes by considering the above factors, and then showed it in the form of mathematical expressions. To make high-precision for each unknown node, the number of the optimal anchor nodes and the position of the anchor nodes are all different based on the different localization, so as to make the best average localization accuracy of the entire networks.Finally, the improved algorithm was simulated on Matlab with the different shapes and different sizes of sensing area, the different parameters of different anchor nodes proportion and the different communication radius of the sensor nodes. Then we compared and analyzed the results with the DV-Hop and the similar improved algorithms. The results of the comparison show that the novel localization algorithm efficiently improves the localization accuracy of 10% with short communication radius, owns good robustness.